Wavelet-based denoising using subband dependent threshold for ECG signals

This paper employs a wavelet-based denoising technique for the recovery of signal contaminated by white additive Gaussian noise and investigates the noise free reconstruction property of universal threshold. A new thresholding procedure is proposed, called subband adaptive. The parameters of this procedure are chosen by difference in mean method. Simulations are carried out in MATLAB using various ECG signals. The results show that the proposed thresholding technique outperforms the existing thresholding techniques.